Glossary

A semver-style string (vMAJOR.MINOR.PATCH) stamped on every published record. A PATCH bump is a recompute on unchanged math; a MINOR bump tightens a calibration; a MAJOR bump changes the formula or factor set. Within one MAJOR, no field is removed or retyped.

More: Versioning policy

The label attached to a Fun Score's bucket: Exceptional, Very Fun, Fun, Some Fun or Quiet. Bands are derived from the score and are meant for readers; integrators should key off the numeric score.

The mapping from the raw composite (a weighted blend of positive and negative pillar scores) onto the readable 0–100 Fun Score. The calibration anchors the lower and upper score endpoints to the distribution of raw composites across all cities, so a Fun Score of 50 always means the same thing relative to the national field.

More: Methodology

A factor in the registry whose data fetcher is wired but not yet running across every city. Today none of the 23 factors are deferred — all 23 are live and ranked for every city. When a factor is deferred, the recompute fills the pillar with a §6.5 gap-fill placeholder — every city gets the median value of 50 — so the composite Fun Score is unaffected (a constant offset cancels across the ranking) but a leaderboard ranking cities by that one factor would be a single tie. Deferred factors are flagged "coming soon" on the methodology table and rendered as stub pages on /best/<slug>. (If a future fetcher rotation re-introduces a deferred factor, the name will appear here automatically.)

More: How the methodology handles missing data

The score a negative pillar is presented under on its leaderboard, computed as 100 − the raw negative pillar score so that higher means better on the named good thing. A "safety score" is 100 − crime, a "clean-air score" is 100 − air quality, and so on. 7 of the boards are flipped this way — "safest cities", "lightest traffic", "cleanest air" — so a calmer, lower-friction city rises to the top. The flip is pure presentation; the underlying pillar and the composite Fun Score are unchanged.

More: Browse the leaderboards

One measurable input to the score, e.g. live music, food, traffic. Each factor is itself measured from one or more sources and turned into a 0–100 pillar score before being weighted into the composite. There are 23 factors total: 16 positive and 7 negative.

A single number from 0 to 100 that estimates how much there is to do in a city — the headline metric of the index. Built from 23 factors covering food, music, nightlife, the outdoors and more, set against frictions like crime, traffic and disaster risk. In the API and JSON bundles this value is carried in the universal_score field.

More: How it is calculated

A factor whose published 0-100 score sits on a raw measurement of zero — the underlying source returned nothing for this city, so the percentile path lands the city at the bottom of the distribution by default. The composite Fun Score is unaffected (the score still contributes at its published value), but a row marked "Limited data" on a city page is honest about the input: the score reflects an absence of data rather than a low measurement. Distinct from a deferred pillar, where the whole factor is on hold across every city; Limited data is a per-city, per-pillar disclosure.

More: Compare: deferred pillar

A pillar whose higher value makes a city less fun, e.g. crime, traffic, disaster risk. 7 pillars are negative. Their weighted composite is multiplied by a drag coefficient and subtracted from the positive composite before calibration.

A pillar score expressed as its position in the national distribution, not as a raw count. A score of 80 means the city sits at the 80th percentile for that factor across the full index. Percentiles make cities of different sizes directly comparable.

The 0–100 score for a single factor. Each city has 23 pillar scores; they roll up by weight into the positive and negative composites, which calibrate to the Fun Score.

The relative importance assigned to a pillar in the composite. Weights are public and fixed in the published methodology to keep cross-city comparisons fair. The personalize page lets a reader rewrite weights for themselves without changing the published index.

More: Personalize

A factor for which a city had no genuine observation at all, so the recompute substitutes the §6.5 gap-fill value — the median across cities that do have data — which lands the city near the 50th percentile. A cell marked "Placeholder" — on a city page, a leaderboard, or the compare table — is honest that the score is an estimate standing in for a missing measurement, not a measured midpoint. The composite Fun Score is unaffected (a near-median input barely moves the ranking). This is distinct from "Limited data", where a real measurement of zero lands the city at the bottom of the distribution: gap-fill means no observation existed, limited data means the observation was zero. A city, pillar cell is one, the other, or neither — never both.

More: Compare: limited data

A pillar whose higher value makes a city more fun, e.g. live music, food, outdoors. 16 pillars are positive. Their weighted composite is the upside term in the formula.

positive − 0.4 × negative, before the 0–100 calibration. This is the pre-calibration number you can reproduce from the per-city pillar scores; the published Fun Score is the calibrated form of it.

Scores are recomputed daily, on a published schedule. The underlying signals refresh on a tiered calendar — weekly for events, monthly for crime and news, and quarterly to annual for slow government feeds like air quality, weather and census data — so each score keeps tracking a city without churning on noise.

Illustrative bundles published while the live data pipeline is being finalised. Marked on every city page and tagged with a -sample suffix in algorithm_version. Do not integrate against a -sample version in production.

The nearest-neighbour cities by full-factor profile, computed from the pillar scores. Two cities are similar if they score alike across the whole factor set, not just on the headline Fun Score. Published as a short list of slugs on each city detail bundle.

A city's stable URL identifier — lower-case, hyphenated, with the two-letter state suffix, e.g. austin-tx. Use the slug to build the per-city URL on the site (/city/{slug}) or the per-city API endpoint (/v1/city/{slug}).

When two or more cities share the same Fun Score (or the same pillar score), they share a rank. The site uses standard competition ranking — 1 plus the count of cities with a strictly higher score — and tags shared ranks with a T prefix (T#1, T-1st). Because the final calibration clamps and rounds to an integer in 0–100, multi-way ties at the ceiling and floor are normal, not a rendering bug.

More: How a score is built